Speech and EEG Signal Analysis

This research work proposed the Voice Signal and EEG (Electroencephalography) signal based multimodal biometric system based on feature level fusion and Dynamic Time Warping, Manhattan Distance and Euclidean Distance. There are two databases are used for experimentation in this thesis, a first Unimodal biometric system developed using Voice Signal and EEG Signal modalities. The Voice signals and EEG signals each evaluate the performance of the dataset and feature level algorithm are used. Feature level fusion approach is used for KVKRG Voice signal and KVKRG EEG signal modalities. The distance classification Euclidean distance measure techniques are used on feature fusion metrics and performance measured. The Dynamic Time Warping algorithm is used for the classification of Fusion metric and then performance measure. And also the Manhattan distance measure is applied to the feature level fusion of Voice and EEG modalities to classify the feature and then performance measured.